We examine the influences of diversity on level and growth rate of GDP at the state and province level in the United States and Canada. Although cross-country studies typically claim that diversity reduces productivity, recent studies with US datasets argue that cultural, linguistic and/or racial diversity contributes positively to aggregate economic performance. The notion is that regions that promote diversity stimulate innovation by building complementarities in knowledge, skills and experiences. Employing a cross-sectional dataset for 48 contiguous states in the US and 10 provinces in Canada, our work measures the impact of ‘social divergence’ - defined as economic effects of barriers to communication created through differences in language, culture, race, ethnicity or religion - on per capita output. We include three measures of diversity, using fractionalisation indices for linguistic, religious, and cultural differences across the states/provinces; and a set of control variables: educational attainment, urbanization, population density, percentage of working age population, and an index of economic freedom. To test the social divergence hypotheses, an interaction term, with a proxy for low level of English fluency, is also included in the OLS estimation. Based on an exhaustive set of robustness tests, we obtain two robust and statistically significant results: one, the estimated coefficients on diversity, variously measured, have a positive impact on per capita output across states and provinces; two, the coefficient of interaction variable is negative, implying that barriers to communication mitigate the positive effects of diversity. The robustness of the OLS results is reinforced when Instrumental Variable Estimation is employed for the potential endogenous variables. Our findings contribute to current policy debates around ‘state-sponsored’ multiculturalism in Canada and provide fresh insights on the use of social capital as a public policy tool in all settler societies. We intend to carry out further research at the city/county level.[Show full abstract]
![[PDF]](/themes/LURA/images/mimes/b_pdf.png "PDF file"){kind=small}
